Conventionally, in a motion control system required to perform a highly accurate motion operation using a number of axes, such as a machine tool, a semiconductor manufacturing apparatus, or an industrial robot, it has been necessary to mutually synchronize a motion controller and the axes controlled by the motion controller during operation. As a technique used in a conventional motion control system, there is a technique for making it possible to output a command to a servo drive in a desired control cycle even if the command is a multi-axis interpolation command that cannot be processed in the desired control cycle by a single motion controller, in which the same interpolation commands are inputted from a number of motion controllers, servo position commands or servo torque commands for a number of servo drives are produced while their production timings are shifted little by little, and these commands are transmitted to the servo drives. Because timings from issuance to execution of the commands are not constant for each axis, it is possible to match operations for the axes by shifting the issuance timings of the commands.
As a conventional technique, there is known a network system in which one motion controller and a number of servo drives are connected by a servo network. In the servo network, as in IEC (International Electrotechnical Commission) 61491 or IEEE (Institute of Electrical and Electronics Engineers) 1394, motion command data are transmitted cyclically and at the same timing from the motion controller to the plural servo drives using a network having a synchronous communication function of transmitting a given number of data sets in each cycle set in advance. Further, the network having such a synchronous communication function also has a function of causing, in each cycle, an activation interrupt to nodes connected to the network, and uses this function to cause the servo drives to execute servo control processing in the servo drives cyclically and at the same timing thereby to realize synchronization of the motion controller and a plurality of axes controlled by the motion controller.
Patent Literature 1 discloses an invention for solving a problem that a network suffers reduction of accuracy of a synchronization protocol to synchronize times among apparatuses due to fluctuation in a network delay. In this invention, when two apparatuses, which perform time synchronization, exchange timing packets to thereby calculate a time offset and adjust a local clock, a network delay related to the timing packets is measured and, when the measured network delay exceeds a threshold, the timing packets are discarded. Consequently, in time synchronization processing of the network, a markedly large network delay is not used. Therefore, it is possible to prevent the accuracy of the time synchronization protocol from being deteriorated.